1. Field of the Invention
The present invention relates to a method and a device for image segmentation.
2. Description of the Prior Art
Touch displays are widely used in may electronic devices. One approach is to define a two dimensional (2D) touch area on a touch display with the use of a touch sensing panel, so as to obtain sensing information through scanning of horizontal and vertical axes on the touch panel in order to determine the touch or proximity of an external object such as a finger on the touch panel. A capacitive touch display is provided by U.S. Pat. No. 4,639,720, for example.
Sensing information can be converted into a plurality of continuous signal values by an analog-to-digital converter (ADC). The location of the touch or proximity of an external object on the touch panel can be determined by comparing the changes in signal values before and after the touch of the external object.
In general, a controller controlling the touch panel will first obtain the sensing information when there is no touch or proximity of any external object as the baseline value. For example, in a capacitive touch panel, each conductive strip corresponds to a respective baseline value. The controller compares subsequent sensing information with the baseline values to determine if there is a touch or proximity of an external object, and to further determine the location of the external object. For example, when there is no touch or proximity of an external object, the subsequent sensing information with respect to the baseline value will be a zero value or approaching a zero value. Whether or not there is a touch or proximity of an external object can be determined by checking whether the sensing information with respect to the baseline value is a zero value or approaching a zero value.
As shown in FIG. 1A, when an external object 12 (e.g. a finger) touches or approaches a sensing device 120 of a touch display 10, sensing information of sensors 140 on an axis (e.g. X-axis) are converted into signal values as shown in FIG. 1B, corresponding to the shape of a finger. The signal values displays a waveform or a finger profile. The location of the peak 14 on the finger profile represents the location of the touch or proximity of the finger.
Typically, 2D sensing information consists of a plurality of 1D sensing information. In other words, performing detections a number of times on the plurality of sensors is needed to generate the 2D sensing information, which will take a consider amount of time. When a high sampling frequency of touch locations is required, it becomes critical to reduce the time taken to generate the 2D sensing information. However, a large proportion of the 2D sensing information is irrelevant to touches, so continuously performing detections irrelevant to touches is time and power consuming. Thus, how to save time and power is a technically critical.
From the above it is clear that prior art still has shortcomings. In order to solve these problems, efforts have long been made in vain, while ordinary products and methods offering no appropriate structures and methods. Thus, there is a need in the industry for a novel technique that solves these problems.